Web Configuration
Now that the hardware has been completely installed, it's time to configure the Bullet to act as a WiFi Access Point.
Configuration is performed using your web browser by typing in https://192.168.1.20. You'll get a warning pop up stating that the device has presented an invalid SSL certificate due to the certificate being self-signed and not issued by a CA. This is normal so just hit accept/continue. You'll then need to log in with the default username and password of "ubnt".
Configure the Wireless tab
The first step is to configure the Wireless tab so the Bullet acts as a WiFi Access Point. Change the settings according to the below picture. Be sure to set the device up with WPA2 security and an appropriate WiFi password here. You can also use this tab to adjust the power output should you wish to cover a smaller area. The default channel bandwidth is 40MHz, but some older devices such as wireless printers and mobile phones cannot operate on 40MHz, so we've changed it to 20MHz in this setup. WDS mode is not required in this instance.
Turn off airMAX
Navigate to the left-most tab (with the Ubiquiti logo) and uncheck the airMAX checkbox to disable Ubiquiti's proprietary airMAX protocol (a special MIMO TDMA multiplexing technique that allows communication between two or more Ubiquiti devices). You won't be able to untick the box until the Wireless tab has been completely configured. AirMax has the ability to increase wireless channel usability from 30% (on 802.11n) to as high as 90%.
Configure Network Settings
In this case our Dovado router is located at 192.168.0.1 with a subnet mask of 255.255.255.0, meaning all our devices are restricted to the 192.168.0.XXX subnet. This means that we've had to change the default 192.168.1.20 to 192.168.0.20. As we want the Dovado to continue controlling the network we're setting the Bullet to Bridge mode, and then assigning the Default Gateway and DNS IP addresses to the address of the router. We've also switched off the Dovado's WLAN just to free up some spectrum as we've already got a collocated ADSL WLAN which wouldn't have left much space for the Bullet.
Performance Comparison
To assess performance we're going to test the Bullet M2 versus the more substantial Rocket M2 setup. Our location in a mixed commercial/residential area is an ideal proving ground to demonstrate resistance against co-channel interference (ie collocated AP's, residential WiFi, etc), multipath propagation through dense urban environment consisting of both trees and buildings.
Both the Bullet and Rocket were mounted (at separate times) on the same mast which was positioned indoors at the roller-door entrance of our warehouse. A simple walk-test was conducted using a simple program called inSSIDer which provided dBm readings of both our test AP and neighbouring (interfering) AP's, along with good old fashioned command prompt to provide some basic packet loss statistics. Performing a ping test is a great simple way to test link conditions - ICMP does not have a guaranteed delivery mechanism so deteriorating radio conditions are instantly obvious. Ideally a more comprehensive WiFi test should be conducted withJPerf/IPerf but here I'd prefer to demonstrate some basic stats that can be collected on your home network without requiring any extra tools.
Bullet Range
The Bullet M2 had a stable range of about 150m through a typical suburban non-LOS environment. Range extended out to about 200m if higher packet loss was tolerable (11% loss @ 194m) - this sort of loss might be acceptable for a simple telemetry station, but not for bandwidth intensive applications. With a range about 400% greater than a normal WiFi access point this is quite a good result, especially in the presence of 24 detectable and 6 overlapping access points.
Rocket M2 Range
Before we begin it's important to note that this may be a somewhat unfair comparison - the smallest available Rocket M2 base station antenna is 10dBi. To compensate for the higher gain we reduced the dBm output by 4dB, however as the radiation pattern differs this should not be overlooked.
Walking out to the 194m mark where we experienced an 11% packet loss previously, this time not a single packet was lost. It took as walking out to 325m before similar levels were experienced, with a usable range of about 275m. Both maximum and average latency were also significantly reduced - by almost half that of the Bullet.
Conclusions
While this simple demonstration should in no way be considered (and was never intended to be) empirical in nature, it's quite clear that the Ubiquiti Rocket outperformed the Bullet in both range and reliability. While Rocket does not employ smart antenna techniques such as beamforming, multiple antenna technology clearly provided an advantage over a single antenna WiFi radio, not surprising in an environment relying entirely on radio scatter. Ubiquiti Bullet is an ideal compact form factor WiFi solution, suitable for a much more diverse range of applications and environments. However for maximum range and performance, the choice is still Ubiquiti Rocket.
